Image courtesy of Solen Feyissa | Unsplash

This article is sponsored by Parsons Behle & Latimer.

The meteoric rise in the use of artificial intelligence (AI) in today’s business setting is unprecedented. In addition, the risks associated with AI and its output give rise to heightened concerns of liability by businesses using or considering using these emerging technologies.

AI is “trained” on content that could include a third party’s intellectual property (IP) and may have been “scraped” from third-party websites or obtained from other sources. The output generated by AI may contain or incorporate that third party’s IP. If proper licenses and rights are not in place between the AI provider and the content owner, it is possible the AI and its output could infringe the IP rights of the content owner, which often involves the copyright in that content. This in turn could lead to potential legal claims for damages from the content owner for such unauthorized usage. If this occurs, customers of the AI provider are often dragged into these legal proceedings and exposed to potentially significant liabilities.

Customers who obtain software, services or technology from providers often enter into contracts with the provider that contain certain protections against third-party infringement claims. These often come in the form of indemnification. Such an indemnification clause would typically be targeted at holding the provider responsible should the software, services or technology infringe the IP of a third party who, in turn, then files a legal claim against the customer by virtue of the customer’s possession, access or use of the IP of that third party.

Early in the commercial availability of AI, many AI providers distanced themselves from potential liability inherent in their AI offerings in their standard form contracts by inserting broad-scope disclaimers of liability associated with that AI. Today, many AI providers are attempting to motivate widespread commercial adoption of their AI products by offering customers indemnification protection should the AI or its output lead to infringement claims against the customer. These AI infringement indemnification clauses must be carefully reviewed and negotiated to ensure that the protections they purport to give are not greatly restricted or even wholly negated by wide-ranging exceptions and restrictions. Further, the indemnification protection provided must be viewed against the limitation of liability provisions in the contract, since even broad indemnification coupled with a relatively low limitation of liability cap may render the AI indemnification protection quite limited or in some cases even illusory.

If your business is using or considering using AI in its operations, carefully consider the indemnification protections your business is, or in some cases isn’t, receiving under the relevant contract with the AI provider. At Parsons Behle & Latimer, we have expertise in dealing with issues in the AI landscape. Let us know how we can help your business in this regard.